Technical Computing from Domain Analysis to Performance Profiling - PowerPoint Presentation

378 views
Uploaded On 2018-09-22

Technical Computing from Domain Analysis to Performance Profiling - PPT Presentation

Phil Pennington Sr Developer Evangelist Microsoft Corporation SESSION CODE WSV325 Required Slide AGENDA TechnicalComputing Parallel Platform Tools Solvers Analysis Technical Computing Microsoft ID: 674775

parallel core tools microsoft core parallel microsoft tools socket computing library threads numa tpl task windows technical vs2010 parallelism

Link:

Copy

Embed:

<iframe width="560" height="315" src="https://www.docslides.com/embed/674775" frameborder="0" allowfullscreen></iframe>

Download Presentation from below link

Download Presentation The PPT/PDF document "Technical Computing from Domain Analysis..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.

Presentation Transcript

Slide1

Technical Computing from Domain Analysis to Performance Profiling

Phil PenningtonSr. Developer EvangelistMicrosoft Corporation

SESSION CODE: WSV325

Required SlideSlide2

AGENDA

Technical_Computing = Parallel (Platform + Tools + Solvers + Analysis);

Technical Computing @ MicrosoftParallel Tools in Visual StudioThinking ParallelUsing TPL and C# Slide3

An Example

Monte Carlo Approximation of Pi

S = Area of square S = (2*r) * (2*r) = 4*r *rS = 4*r*r

DEMO

C = Area of Circle

C = Pi*r*r

Pi = 4 * (Area of Circle / Area of Square)Slide4

DEMO

S = 4*r*r

C =

Pi*r*r

Pi = 4

*(C/S)

An Example

Monte Carlo Approximation of Pi

For each Point (P),

d(P)

= SQRT

((x * x) + (y * y

))

if (d < r) then (

x,y

) in CSlide5

Why Parallel?Slide6

Windows and Logical Processors

Before Win7/R2, the maximum number of Logical Processors (LPs) was dictated by processor integral word size

LP state (e.g. idle, affinity) represented in word-sized bitmask32-bit Windows: 32 LPs64-bit Windows: 64 LPs

32-bit Idle Processor Mask

Idle

BusySlide7

Windows Organizes Many-Cores via GROUP

New with Windows 7

and R2

GROUP

NUMA NODE

Socket

Core

NUMA = Non-Uniform Memory Access

LP = Logical ProcessorSlide8

Processor Groups

Example: 2 Groups, 4 Nodes, 8 Sockets, 32 Cores, 4 LPs/Core = 128 LPs

Group

NUMA Node

Socket

Core

Socket

Core

NUMA Node

Socket

Core

Socket

Core

Group

NUMA Node

Socket

Core

Socket

Core

NUMA Node

Socket

Core

Socket

Core

LPSlide9

CPU0

CPU1

CPU2

CPU3

Static Scheduling

Load-Balancing Task Scheduler

Dynamic

scheduling improves performance by distributing work efficiently

at runtime.

CPU0

CPU1

CPU2

CPU3

Dynamic

SchedulingSlide10

Your Scheduler

Logic

Reason:

Yield

Wait

Reason:

Yield

User Mode Scheduling

Architectural Perspective

Application

Kernel

Scheduler Threads

CPU 1

CPU 2

Blocked Worker Threads

UMS Scheduler’s

eady List

UMS Completion List

Reason:

Created

Reason:

BlockedSlide11

The Platform

Topology

DEMOSlide12

AGENDA

Technical_Computing = Parallel (Platform + Tools + Solvers + Analysis);

Technical Computing @ MicrosoftParallel Tools in Visual StudioThinking ParallelUsing TPL and C# Slide13

Tasks in .NET and C++

.NET 4.0

Parallel.For(x, y, λ)Parallel.ForEach(IEnum

)

Parallel.Invoke

(

, λ)TaskTask.Factory.StartNew

(

)

ThreadPool-basedVisual C++ 10parallel_for(x, y, step, λ);parallel_for_each(it, λ)parallel_invoke(λ, λ)task_group / task_handletask_group::run (λ)Native concurrency runtime(and many overloads for the above)Slide14

Tools

Programming Models – Structured Parallelism

.NET Parallel Extensions

.NET Runtime

Visual Studio 2010, .NET Developer

Tools, Programming Models, Runtimes

Parallel LINQ

(PLINQ)

Resource Manager

Task Scheduler

Managed Library

Threads Pools

Data Structures

Tools

Parallel

Debugger

Parallel

Profiler

Task Parallel

LibrarySlide15

Tools

Programming Models – Structured Parallelism

C++ Concurrency Runtime

Operating System

Visual Studio 2010, C++ Developer

Tools, Programming Models, Runtimes

Parallel Pattern Library

Resource Manager

Task Scheduler

Native Library

Key:

Threads

Data Structures

Tools

Parallel Debugger

Parallel

Profiler

Agents

Library

Win7/R2: UMS ThreadsSlide16

Capabilities Comparison (1)

VS2010

PCP Technologies

Capability

.NET4

TPL

C++

ConcRT

OpenMP

PLINQ

MPI

Threads/Thread-Pools

Task Parallelism

N+

Y-

Data

Parallelism

N+

Parallel Patterns

N+

Fine-grained

Parallelism (loops)

Y-

Work-Item

Partitioning

Y-

N+

Dynamic Scheduling

N+

NSlide17

Capabilities Comparison (2)

VS2010

PCP Technologies

Capability

.NET4

TPL

C++

ConcRT

OpenMP

PLINQ

MPI

Threads/Thread-Pools

Affinity

Y-

Concurrent

Data Structures

N+

Scalable

Memory Allocator

Optimized

I/O Capability

User-Mode

Sync Primitives

Automatically

Collates Results

NSlide18

The Tools

Libraries

Languages Debuggers

Profilers

DEMOSlide19

AGENDA

Technical_Computing = Parallel (Platform + Tools + Solvers + Analysis);

Technical Computing @ MicrosoftParallel Tools in Visual StudioThinking Parallel

Using TPL and C# Slide20

Thinking Parallel - “Control” vs. “Data”

ParallelismControl Parallelism

Parallel.For (0, size, (i) => { Console.WriteLine(i);

});

Data Parallelism

IEnumerable

<int

> numbers =

Enumerable.Range

(2, 100-3);

var parallelQuery = from n in numbers.AsParallel() where Enumerable.Range(2, (int)Math.Sqrt(n)).All(i => n % i > 0) select n; int[] primes = parallelQuery.ToArray(); Slide21

Thinking Parallel – How to Schedule my Tasks?Slide22

Thinking Parallel – How to Partition my Data?

Several partitioning schemes built-in

Chunk

Works with any

IEnumerable

<T>

Single enumerator shared; chunks handed out on-demand

Range

Works only with

IList

<T>